Skip to content
Tutorials Center
My learning

Topics

Lesson 13 of 38 · Wavefunctions and Probability

Momentum Space and the Fourier Transform

Switch from position to momentum description with the Fourier transform, compute φ(k) numerically and see the trade-off between width in x and width in k.

The wavefunction ψ(x) tells you about position. But the same state can be described equally well by a function of momentum, φ(k) (or φ(p)). The two are linked by the Fourier transform. Understanding this link explains the uncertainty principle at a glance: a function cannot be narrow in both descriptions at once.

The transform pair

With wavenumber k (momentum p = ħk), the standard quantum convention is

φ(k) = (1/√(2π)) ∫ ψ(x) e^{-ikx} dx

ψ(x) = (1/√(2π)) ∫ φ(k) e^{ikx} dk

The factor 1/√(2π) in both directions makes the transform unitary: it preserves normalization (Plancherel's theorem):

∫ |ψ(x)|² dx = ∫ |φ(k)|² dk

If ψ is normalized, so is φ. By the Born rule, |φ(k)|² dk is the probability that a momentum measurement gives a wavenumber between k and k + dk, so the momentum probability density is |φ(k)|² (times 1/ħ in p units).

In momentum space, position is the "derivative" operator x̂ = i d/dk and momentum is simply multiplication: p̂ = ħk. Both pictures contain the same information.

The Gaussian is special

The Fourier transform of a Gaussian is a Gaussian. If

ψ(x) = (2πσ²)^(-1/4) e^{-x²/(4σ²)}

then φ(k) = (2σ²/π)^(1/4) e^{-σ² k²}, so |φ|² is a normal distribution in k with standard deviation σ_k = 1/(2σ). Multiply by ħ: Δp = ħ/(2σ), matching the previous lesson. Narrow in x means wide in k, and vice versa.

Two descriptions

ψ(x) Where is it?
Fourier Unitary transform
φ(k) How fast is it going?
1 / 2

Computing φ(k) with the FFT

On a grid xⱼ = x₀ + j dx, the transform integral becomes a sum, which numpy.fft.fft computes quickly. With the grid frequencies k = 2π fftfreq(N, dx),

φ(k) ≈ (dx/√(2π)) e^{-i k x₀} · fft(ψ)

The phase factor e^{-ik x₀} accounts for the grid starting at x₀; it never changes |φ|². Use np.fft.fftshift to put k in increasing order for plotting.

Python Fourier transform of a Gaussian: check widths

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The norm is 1 in both descriptions, and the numeric σ_k agrees with 1/(2σ).

A sharp edge needs many momenta

A function with a sudden jump needs very high-frequency components. The rectangular pulse of width a has

φ(k) = √(a/(2π)) · sinc(ka/2), with sinc(y) = sin(y)/y.

Its momentum distribution has a main peak of half-width about 2π/a and slowly decaying side lobes (1/k² in probability). In practice, a hard-walled box is a poor localizer in momentum space: ⟨p²⟩ of a function with a discontinuity is infinite. The smooth box ground state sin(πx/L) has finite ⟨p²⟩.

Python Momentum distribution of a rectangular pulse

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The two agree to within about 0.004 (the sharp edges are only approximated on a grid), and the zeros occur at k = 2πn/a.

Momentum-space probabilities in practice

To get the probability that the momentum lies in a range, integrate |φ|² over that range; this is how electron-diffraction and spectroscopy data are often interpreted. Note that a particle in a bound state (for example, an electron in an atom) does not have a single momentum; its momentum distribution has a width inversely related to its size.

Momentum "p" and wavenumber "k" differ by a factor ħ. Some textbooks write φ(p) = (1/√(2πħ)) ∫ ψ e^{-ipx/ħ} dx. Both forms are equivalent; we use k because it keeps formulas tidy.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. What does Plancherel's theorem say?

    The unitary Fourier transform preserves the norm, so a normalized ψ gives a normalized φ.
  2. 2. The Fourier transform of a Gaussian in x is

    Gaussians are mapped to Gaussians, with reciprocal widths.
  3. 3. If you squeeze a wavefunction to be narrower in x, its momentum distribution

    Narrow in x requires a broad superposition of wavenumbers.
  4. 4. In momentum space, the momentum operator is

    Acting on φ(k), p̂ simply multiplies by ħk.

Summary

  • φ(k) and ψ(x) are Fourier transforms of each other and carry the same information.
  • |φ(k)|² is the momentum probability density; norms agree (Plancherel).
  • A Gaussian of width σ in x has width 1/(2σ) in k.
  • Sharp features in x need broad momentum content; compute transforms numerically with the FFT.

Next: The Time-Dependent Schrödinger Equation